In my research project on deforestation I use the indigenous and protected lands in the legal Amazon from TerraBrasilis:
https://t.co/fq6kwODAi5
I need the year that each of these was established. Any ideas where to get it?
#deforestation#brazil#amazonrainforest
Inspecting the attention scores allows to better understand what features are important for recognizing each driver.
It also shows that the model implicitly learns to ignore cloudy images (scores equal to 0). No need for any cloud masking or filtering! ☔️
(3/3)
Why are we losing tropical forests? 🌴
Together with @devistuia, @MarcCoru, @JanDirkWegner1, and others, we used satellite images to map the drivers of forest loss across tropics
Read the full paper:
https://t.co/e8as0nXYU0…
(1/3)
To do this, we developed a deep learning model that classifies drivers from time series of Sentinel-2 images.
While making a prediction, it also assigns an attention score to each image of the time series, indicating their relative importance.
(2/3)
PhD position on NLP+Earth observation at Inria Montpellier: https://t.co/AKCRGslfCt
In the context of the GEO-ReSeT project (Inria and Université de Paris Cité), we are looking for someone to help us build a language-aware Earth observation foundation model.
@mohitanand Hi Mohit, I'm not at AGU but this sounds really cool and I would like to learn more. Is there a recording or a publication I can check out?
Thank you!
@RicardoDalagnol Thanks Ricardo. It seems we are working on similar topics :) Good luck with the annotations and looking forward to your large-scale maps!
@CTrees_org Whether this is associated with a change in land use may only become apparent much later (e.g., if forest is cut down and then let to regrow, there is no land use change but I believe the alerts would detect it)
@CTrees_org Hi, can you please explain how you detect land use change? If I understand correctly, the radar alerts detect change in land *cover* (there is forest -> there is no forest).
📢 We are hiring! 📢 Interested in sustainable mobility, data science, spatio-temporal modelling and big data? Come work with me and our talented GIST Lab team @AaltoUniversity. We have 2 positions open related to a @SuomenAkatemia project starting next fall. #gischat Details: 👇
Want to learn about meta-learning & few-shot learning?
All of the latest lecture videos for Stanford CS330 are now online!
https://t.co/PqNK2yqQ16
New topics in Fall '22 include:
- self-supervised pre-training
- large scale meta-optimization
- domain adaptation & generalization
Job alert! 🚨
We are looking for a post-doc at Inria Montpellier to work on Species Distribution Models with Remote Sensing data by combining Deep Learning and Bayesian models. Join us! 🛰️🗺️🌲🌳🌴